AI Research
50 articles in this category
Diffusion Research: Drug Discovery Outshines Image Generation
AI's diffusion model research is making bigger waves in drug discovery than image generation, tackling complex molecular interactions with physics-informed AI.
Google Boosts AI: Faster Images, Video Tools
Google launches Nano Banana 2 Lite for fast, cheap image generation and makes Gemini Omni Flash available for video editing.
Memora: Microsoft's AI Memory Upgrade
Microsoft's Memora AI memory system revolutionizes long-term AI interactions by balancing detailed recall with efficient retrieval, outperforming existing solutions.
Meta's Nishant Gupta on Deterministic AI Infrastructure
Nishant Gupta from Meta discusses the critical need for deterministic infrastructure to reliably run non-deterministic AI agents, highlighting the shift from model-centric to systems-centric development.
RL Agent Automates ETL Pipeline Failure Remediation
Anna Marie Benzon presents an RL agent designed to automate ETL pipeline failure detection and remediation, significantly reducing recovery time and enhancing system reliability.
OpenAI Weighs 2027 IPO Amid Market Volatility
OpenAI is reportedly considering a 2027 IPO, navigating market volatility and aiming to solidify its position in the rapidly evolving AI landscape.
Benchmarks Fail Modern AI, Says OpenAI Scientist
OpenAI's Noam Brown discusses why traditional benchmarks fail modern AI, emphasizing the need for new evaluation methods that account for computational budgets and model capabilities.
OpenAI's Mark Chen on AGI, Scaling Laws, and Evals
OpenAI's Chief of Research, Mark Chen, shares insights on the path to AGI, the impact of scaling laws, and the importance of robust evaluations for AI safety.
Raymond Weitekamp on Recursive Coding Agents
Raymond Weitekamp of OpenProse discusses recursive coding agents, exploring how AI systems can autonomously generate and refine their own code.
Meta's Nishant Gupta on Evaluating Agentic AI Systems
Nishant Gupta from Meta's Superintelligence Labs discusses the shift from accuracy-based evaluation to reliability-focused methods for agentic AI systems.
OpenData Pipeline Elevates Agentic AI
The OpenThoughts-Agent project introduces an open data pipeline that significantly enhances generalization for agentic language models, outperforming existing benchmarks.
AI Models Storms with Unprecedented Accuracy
AI models are achieving surprising accuracy in predicting mega storms, outperforming traditional methods and offering crucial insights into future weather patterns.
Engram's AI: Memory and Continual Learning
Engram's Dan Biderman and Jessy Lin discuss the critical role of memory and continual learning in AI, aiming to overcome catastrophic forgetting.
AI Security Post-Codex & Claude: Kolter & Fredrikson
AI security experts Zico Kolter & Matt Fredrikson discuss the challenges posed by models like Codex & Claude, and Gray Swan's approach to securing AI.
FlashRT: Execution State for Latency-First AI
FlashRT revolutionizes on-device AI serving with execution-state capsules, enabling sub-millisecond state restoration and significant TTFT speedups for latency-critical applications.
PDE Solutions Get Analytical
Agentic Symbolic Search (ASYS) automates the discovery of analytical forms for PDE solutions, bridging computation and mathematical insight.
Anthropic's Co-founder on AI Research at the Frontier
Anthropic's Co-founder and Top Economist discuss the frontier of AI research, covering economics, safety, and future implications.
Hybrid AI Models Get Orthogonal
OrthoReg, a novel regularization method, ensures clear separation between symbolic and neural components in hybrid dynamical systems, boosting interpretability and generalization.
Autonomous Agents Streamline Data Integration
Data Intelligence Agents (DIA) system revolutionizes data integration by using autonomous coding agents to generate, execute, and validate concrete artifacts, achieving state-of-the-art results.
OneCanvas: Unified 3D Scene Representation
OneCanvas revolutionizes 3D scene understanding in VLMs by projecting multi-view features onto a unified equirectangular canvas, enabling efficient situated reasoning and SOTA performance.
LoopWM: A New Scaling Axis for World Models
Looped World Models (LoopWM) redefine world simulation with iterative refinement, achieving 100x parameter efficiency and establishing latent depth as a new scaling axis.
WEQA: Bridging LLMs and Wearable Health Data
WEQA, a novel agent framework, unifies LLM reasoning with specialized tools for wearable health data, achieving 24% higher accuracy and expert-validated clinical soundness.
UK trials AI for faster house planning
UK government partners with Google DeepMind on an AI prototype to drastically cut house planning application times.
Phase Dominance in AI Image Recognition
AI image classifiers exhibit a striking phase dominance for identity encoding, mirroring human vision principles, with architectural differences shaping its expression.
TokenPilot: Reining in LLM Context Costs
TokenPilot offers a dual-granularity context management framework, slashing LLM inference costs by up to 87% while preserving performance.
ActiveSAM: Efficient Open-Vocabulary Segmentation
ActiveSAM revolutionizes open-vocabulary semantic segmentation with a training-free framework that dynamically identifies relevant classes, boosting speed and accuracy while enhancing robustness for real-world AI.
Nvidia's Ziv Ilan on Faster Diffusion Models
Nvidia's Ziv Ilan explains how to reduce diffusion model latency using quantization, caching, and distillation, plus the new FastGen library.
Compute Once: Unlocking AI Agent Efficiency
A radical proposal to precompute LLM KV caches, slashing inference costs by up to 50x and enabling a new compute-efficient AI agent paradigm.
HYDRA-X: Unifying Image & Video Tokenization
HYDRA-X, a novel Vision Transformer-based UMM, unifies image and video tokenization, enhancing editing consistency and performance through causal attention and latent-level manipulation.
Humanoids Learn Self-Other Distinction
Humanoid robots now learn self-other distinction and build predictive self-models from sensory data, enabling better collaboration and task performance in human-robot environments.
AI spots new LOTUSLITE variant
Microsoft's AI agent 'Ire' has identified a new LOTUSLITE malware variant missed by traditional security tools, showcasing AI's prowess in behavioral analysis.
Unlocking Ultra-Long Context for LLMs
MiniMax Sparse Attention breaks the context window barrier for LLMs, enabling millions of tokens with significant compute reduction and practical speedups.
Mana Reimagines Dexterous Robotics
Mana framework reinterprets dexterous robotics as animation, achieving zero-shot sim-to-real transfer for articulated tool manipulation.
From LLM Agents to Scientific Knowledge Graphs
Agents-K1 revolutionizes LLM research agents by creating agent-native scientific knowledge graphs from full papers, enabling deeper scientific reasoning.
5 AI Research Papers Shaping AI's Future
Discover five key AI research papers that reveal the current trajectory and future directions of artificial intelligence development.
Rethinking VLM Token Reduction
Reroute transforms VLM token reduction from irreversible pruning to recoverable routing, improving grounding performance without sacrificing efficiency.
Automating Scientific Discovery
ATLAS, an active learning framework, automates the discovery of interpretable mechanistic models, achieving 5-10x sample efficiency gains.
VLA Models Unlock Decentralized Multi-Robot Teams
CHORUS leverages pretrained VLA models for decentralized multi-robot collaboration, achieving significant performance gains without inference-time communication.
Codex Aids Black Hole Simulation Breakthrough
This video explores how the AI model Codex is revolutionizing the creation of black hole simulations, making previously intractable problems computationally feasible and accelerating astrophysical research.
DeepMind's Kilpatrick on AI Models Eating Harnesses
Google DeepMind's Logan Kilpatrick delves into the AI concept of models "eating the harness," explaining how over-specialization hinders generalization and what can be done to prevent it.
Causal Inference's Counterfactual Blind Spot
Predictive AI models fail on counterfactual couplings. A new world model using semidefinite kernels offers a solution for robust causal inference.
Steering LRMs Beyond Output Degradation
A new probe-based method, FPCG, distinguishes prediction from detection features to enable precise large reasoning models steering with minimal output quality degradation.
LLMs Accelerate FPGA Design
LLMs are now automating complex FPGA accelerator design, reducing time and expertise needed for efficient AI hardware deployment.
DiffusionGemma: Google's AI is 4x Faster
Google DeepMind's DiffusionGemma model offers up to 4x faster text generation, enabling new real-time AI applications.
Google DeepMind Discusses Open Models & AI Ownership
Google DeepMind's Gus Martins and Ian Ballantyne discuss the benefits of open AI models like Gemma for ownership, control, and custom applications.
Topology-Aware Operator Learning
Topological Neural Operators (TNOs) provide a unified framework for operator learning on cell complexes, improving PDE benchmark accuracy by integrating topological structures.
Personalized AI Agents Now Have a Benchmark
A new iOSWorld benchmark reveals AI agents' struggles with personalized, multi-app tasks, highlighting the need for richer context and advanced reasoning capabilities.
Images as the New Reasoning Medium
This paper introduces optical reasoning, enabling images to serve as the primary medium for LLM and MLLM reasoning, achieving higher token efficiency and competitive performance.
Gemini's Audio Stack: From Transcription to Music Generation
Google DeepMind's Thor Schaeff explores Gemini's audio stack, from advanced transcription to music generation with Lyria 3.
Google Rolls Out Gemini 3.5 Live Translate
Google's new Gemini 3.5 Live Translate offers real-time speech-to-speech translation across 70+ languages, enhancing Google Translate and Meet.